Research map / Label-efficient and robust learning / Recognition and long-tail
Patch-NetVLAD: Multi-Scale Fusion of Locally-Global Descriptors for Place Recognition
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Add multi-scale patch fusion rather than use a single patch size (ablation) | Patch-NetVLAD with RANSAC scoring · Mapillary Val. set | R@1: 77.8 -> 79.5 | Helped |
| Replace RANSAC with rapid spatial scoring while retaining one patch size (ablation) | Patch-NetVLAD · Mapillary Val. set | R@1: 77.8 -> 77.2 | Hurt |
| Replace RANSAC with rapid spatial scoring while retaining multiple patch sizes (ablation) | Patch-NetVLAD · Mapillary Val. set | R@1: 79.5 -> 78.2 | Hurt |
| Add multi-scale fusion while using rapid spatial scoring (ablation) | Patch-NetVLAD · Mapillary Val. set | R@1: 77.2 -> 78.2 | Helped |
| Use rapid spatial verification instead of RANSAC (ablation) | Patch-NetVLAD with one patch size · Place recognition datasets | processing speed: 2.9 times faster | Helped |
| Use rapid spatial verification instead of RANSAC (ablation) | Patch-NetVLAD with multiple patch sizes · Place recognition datasets | processing speed: 3.1 times faster | Helped |
| Use rapid spatial scoring instead of RANSAC (ablation) | Patch-NetVLAD with one patch size · Place recognition datasets | recall: 0.6% | Hurt |
| Use rapid spatial scoring instead of RANSAC (ablation) | Patch-NetVLAD with multiple patch sizes · Place recognition datasets | recall: 1.1% | Hurt |
| Increase patch descriptor PCA dimension (ablation) | Patch-NetVLAD · Mapillary validation set | R@1 | Mixed |
| Vary the single patch size away from the intermediate-size peak (ablation) | Single-Patch-NetVLAD · Mapillary validation set | recall | Hurt |
Closest papers by meaning
- SuperVLAD: Compact and Robust Image Descriptors for Visual Place Recognition (NeurIPS 2024 · similarity 0.91)
- Towards Implicit Aggregation: Robust Image Representation for Place Recognition in the Transformer Era (NeurIPS 2025 · similarity 0.89)
- Optimal Transport Aggregation for Visual Place Recognition (CVPR 2024 · similarity 0.89)
- Focus on Local: Finding Reliable Discriminative Regions for Visual Place Recognition (AAAI 2025 · similarity 0.88)
- Towards Seamless Adaptation of Pre-trained Models for Visual Place Recognition (ICLR 2024 · similarity 0.88)
- MutualVPR: A Mutual Learning Framework for Resolving Supervision Inconsistencies via Adaptive Clustering (NeurIPS 2025 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.